Abstract
Research on speech recognition has existed for a long time, but there is very little research on applying deep learning to Vietnamese language speech recognition. In this paper, we solve the Vietnamese speech recognition problem by deep learning speech recognition frameworks including CTC and Joint CTC/Attention combined with encoder architectures Conformer. Experimental results achieved moderate accuracy using over 115 h of training data of VLSP and Vivos. Compared with the other models, the training results show that the Conformer model trained on CTC achieved good results with a WER value of 20%. Training on big data gives remarkable results and is the basis for us to continue improving the model and increasing accuracy in the future.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Li, J.: Recent advances in end-to-end automatic speech recognition. ArXiv arXiv:2111.01690 (2022)
Graves, A., Fernández, S., Gomez, F., Schmidhuber, J.: Connectionist temporal classification: labelling unsegmented sequence data with recurrent neural ’networks. In: ICML - Proceedings of the 23rd International Conference on Machine Learning, vol. 2006, pp. 369–376, New York (2006). https://doi.org/10.1145/1143844.1143891
Chan, W., Jaitly, N., Le, Q., Vinyals, O.: Listen, attend and spell: a neural network for large vocabulary conversational speech recognition. In: IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 4960–4964 (2016). https://doi.org/10.1109/ICASSP.2016.7472621
Graves, A.: Sequence transduction with recurrent neural networks. In: CoRR. (2012). ArXiv arXiv:1211.3711
Kim, S., Hori, T., Watanabe, S.: Joint CTC-attention based end-to-end speech recognition using multi-task learning. In: 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 4835–4839 (2017)
Amodei, D., et al.: Deep speech 2: end-to-end speech recognition in english and mandarin (2015). ArXiv arXiv:1512.02595
Han, W., et al.: ContextNet: improving convolutional neural networks for automatic speech recognition with global context (2020). ArXiv arXiv:2005.03191
Gulati, A., et al.: Conformer: convolution-augmented transformer for speech recognition (2020). ArXiv arXiv:2005.08100
Beckmann, P., Kegler, M., Saltini, H., Cerňak, M.: Speech-VGG: a deep feature extractor for speech processing (2019)
Vaswani, A., et al.: Attention is all you need (2017). ArXiv arXiv:1706.03762
Graves, A., Mohamed, A., Hinton, G.: Speech Recognition with Deep Recurrent Neural Networks, (2013). ArXiv arXiv:1303.5778
Nguyen V.H.: an end-to-end model for vietnamese speech recognition. In: 2019 IEEE-RIVF International Conference on Computing and Communication Technologies (RIVF), pp. 1–6 (2019). https://doi.org/10.1109/RIVF.2019.8713758
Nguyen, Q.M., Nguyen, T.B., Pham, N.P., Nguyen, T.L.: VAIS ASR: building a conversational speech recognition system using language model combination (2019). ArXiv arXiv:1910.05603
Zhang, Q., et al.: Transformer transducer: a streamable speech recognition model with transformer encoders and RNN-T Loss (2020). ArXiv arXiv:2002.02562
Dong, L., Xu, S., Xu, B.: Speech-transformer: a no-recurrence sequence-to-sequence model for speech recognition. In: 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 5884–5888 (2018). https://doi.org/10.1109/ICASSP.2018.8462506
Luong, H., T., Vu, H., Q.: A non-expert Kaldi recipe for Vietnamese speech recognition system. In: Proceedings of the Third International Workshop on Worldwide Language Service Infrastructure and Second Workshop on Open Infrastructures and Analysis Frameworks for Human Language Technologies (WLSI/OIAF4HLT2016), Osaka, Japan, The COLING 2016 Organizing Committee, pp. 51–55 (2016)
Institute VinBigdata Homepage. https://institute.vinbigdata.org/en/events/vinbigdata-shares-100-hour-data-for-the-community. Accessed 5 June 2022
Nguyen, B.: Vietnamese end-to-end speech recognition using wav2vec 2.0. (2021). https://doi.org/10.5281/zenodo.5356039
Thanh, N.T.M., Dung, P.X., Hay, N.N., Bich, L.N., Quy, Đ.X.: Đánh giá các he thong nhan dạng giọng nói tieng Viet (Vais, Viettel, Zalo, Fpt và Google) trong ban tin. In: Tạp Chí Khoa Học Giáo Dục K\(\tilde{\rm {y}}\) Thuat, Vietnamese. vol. 63 (2021)
Nguyen, Q.B., Mai, V.T., Le, Q.T., Dam, B.Q., Do, V.H.: Development of a Vietnamese large vocabulary continuous speech recognition system under noisy conditions. In Proceedings of the Ninth International Symposium on Information and Communication Technology (SoICT 2018), pp. 222–226. Association for Computing Machinery, New York (2018). https://doi.org/10.1145/3287921.3287938
Acknowledgement
We thank the R &D team of EduplaX - WeAi for giving us the time to do a more in-depth study of speech recognition and experiment with the results of this article.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Van Anh Tuan, N., Hoa, N.T.T., Dat, N.T., Tuan, P.M., Truong, D.D., Phuc, D.T. (2022). A Novel Approach for Vietnamese Speech Recognition Using Conformer. In: Dang, T.K., Küng, J., Chung, T.M. (eds) Future Data and Security Engineering. Big Data, Security and Privacy, Smart City and Industry 4.0 Applications. FDSE 2022. Communications in Computer and Information Science, vol 1688. Springer, Singapore. https://doi.org/10.1007/978-981-19-8069-5_53
Download citation
DOI: https://doi.org/10.1007/978-981-19-8069-5_53
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-19-8068-8
Online ISBN: 978-981-19-8069-5
eBook Packages: Computer ScienceComputer Science (R0)